Physiology of Renal Glucose Handling Via SGLT1, SGLT2 and GLUT2

Diabetologia (2018) 61:2087–2097 https://doi.org/10.1007/s00125-018-4656-5

REVIEW

Physiology of renal glucose handling via SGLT1, SGLT2 and GLUT2

Chiara Ghezzi1 & Donald D. F. Loo1 & Ernest M. Wright1

Received: 27 March 2018 /Accepted: 23 April 2018 /Published online: 22 August 2018 # The Author(s) 2018

Abstract The concentration of glucose in plasma is held within narrow limits (4–10 mmol/l), primarily to ensure fuel supply to the brain. Kidneys play a role in glucose homeostasis in the body by ensuring that glucose is not lost in the urine. Three membrane proteins are responsible for glucose reabsorption from the glomerular filtrate in the proximal tubule: sodium−glucose cotransporters SGLT1 and SGLT2, in the apical membrane, and GLUT2, a uniporter in the basolateral membrane. ‘Knockout’ of these transporters in mice and men results in the excretion of filtered glucose in the urine. In humans, intravenous injection of the plant glucoside phlorizin also results in excretion of the full filtered glucose load. This outcome and the finding that, in an animal model, phlorizin reversed the symptoms of diabetes, has stimulated the development and successful introduction of SGLT2 inhibitors, gliflozins, in the treatment of type 2 diabetes mellitus. Here we summarise the current state of our knowledge about the physiology of renal glucose handling and provide background to the development of SGLT2 inhibitors for type 2 diabetes treatment.

Keywords Gliflozins . Glucose . GLUTs . Inhibitors . Kidney . Phlorizin . Proximal tubule . Review . SGLTs . Type 2 diabetes mellitus

Abbreviations Introduction 18 2-FDG 2-Deoxy-2-fluoro-[ F]-D-glucose F-Dapa 4-[18F]fluoro-dapagliflozin Although the kidneys freely filter plasma glucose, none ap- HEK 293 Human embryonic kidney cells pears in the urine. However, in diabetes, glucose may appear

Ki Inhibitor constant in the urine, specifically when the plasma glucose concentra- Km Apparent affinity constant tion is so high that the filtered load exceeds the maximum Me-4FDG Methyl-4-fluoro-[18F]-4-deoxy- capacity for sugar reabsorption. Homer Smith and his col- D-glucopyranoside leagues were the first to quantify glomerular filtration rates OMIM Online Mendelian Inheritance in Man in humans. They showed that all the filtered glucose was nor- PET Positron emission tomography mally reabsorbed and, importantly, they also demonstrated PKA Protein kinase A that reabsorption of total filtered glucose load could be PKC Protein kinase C inhibited by a compound called phlorizin (Fig. 1)[1]. This SGLT Sodium−glucose cotransporter prompted decades of research into the location and mecha- TM Transmembrane helix nism of kidney glucose transport, mostly in animal models.

A brief history of the discovery of renal glucose transporters Early micropuncture experiments on amphibians and rats Electronic supplementary material The online version of this article found that glucose was completely reabsorbed in the proximal (https://doi.org/10.1007/s00125-018-4656-5) contains a slideset of the tubule (Fig. 2). Following the pioneering work of Mo Burg figures for download, which is available to authorised users. who introduced techniques for research on isolated perfused kidney tubules (reviewed in [2]), Barfuss and Schafer [3] * Ernest M. Wright found that the capacity for glucose absorption was tenfold [email protected] larger in early (S2) rabbit proximal tubules than late (S3) tubules. Furthermore, they found that the affinity for glucose 1 Department of Physiology, Geffen School of Medicine at UCLA, Los Angeles, CA 90095-1751, USA was higher in S3 tubules than S2 tubules, and that luminal 2088 Diabetologia (2018) 61:2087–2097

Schultz and Pete Curran largely confirmed and expanded Crane’s hypothesis in the broader context of trans-epithelial transport [4]. Ultimate experimental proof of cotransport came Filtered glucose load Nephron 180 g/day from seminal work by Ulrich Hopfer and colleagues on isolated brush border membrane vesicles [5]. For a summary of Glucose excretion −Pz <0.5 g/day the early history of the cotransport hypothesis, see [6]. The +Pz >179 g/day second step of intestinal glucose absorption, passive exit from

Glucose reabsorption the cell across the basolateral membrane, was thought to occur −Pz >179 g/day by facilitated diffusion. This two-stepped process explains +Pz <1 g/day how glucose absorption can occur against a concentration gradient across the intestinal epithelium (see [4]). An early analysis of two rare genetic disorders of intestinal and renal glucose transport, familial renal glucosuria and glucose–galactose malabsorption, led to the hypothesis that two Fig. 1 A drawing illustrating a cross-sectional view of the human kidney, showing the location of one of the 1 million nephrons in the kidney. The genes are primarily involved in glucose reabsorption in the renal vein and renal artery are shown in black and red, respectively. The kidney, with one of these genes also being expressed in the average daily filtered glucose load, urinary glucose excretion and glucose small intestine (giving rise to glucose–galactose malabsorp- reabsorption in healthy adults are also shown. Phlorizin (Pz) increases tion). Fast-forward to the present day and we now know that glucose excretion to the filtered load and eliminates glucose reabsorption. − This figure is available as part of a downloadable slideset these genes encode the brush border sodium glucose cotransporters (SGLT1 and SGLT2). Moreover, studies of Fanconi–Bickel syndrome, a genetic disease resulting in se- phlorizin blocked absorption in both segments. This led to the vere glucose loss in the urine [7], point to GLUT2 as the concept that the bulk of glucose reabsorption occurred in the kidney basolateral glucose transporter [8, 9]. These genetic early proximal tubule (S1 and S2 segments), and that the later conditions, discussed later, have been instrumental in reveal- segment (S3) ‘mopped up’ the rest. ing the key players in renal glucose transport: SGLT1, SGLT2 Our understanding of the mechanism of sugar transport has and GLUT2. largely rested on the progress stemming from Bob Crane’s In this review, we summarise the physiology of renal 1961 sodium–glucose cotransport hypothesis that glucose SGLTs and GLUTs, and provide rationale for the development transport across the intestinal brush border membrane is uphill of SGLT2 inhibitors to treat type 2 diabetes, as well as and coupled with the downhill movement of sodium. Stan outlining the limitations of this therapeutic approach [10, 11].

ab Glomerulus

S1 Distal

S2 8 Proximal

4 S3 Glucose (mmol/l)

0 02468 Distance of filtrate from glomerulus (mm)

Collecting duct Thin segment

Fig. 2 Glucose reabsorption by the proximal tubule. (a) Schematic rep- segments, and SGLT1 and GLUT2 are responsible for glucose reabsorp- resentation of a single nephron, the functional unit of the kidney. (b) tion in the S3 segment . Adapted from [6], distributed under the terms of Glucose concentration (mmol/l) measured in micropuncture studies as the CC BY 4.0 Attribution License (http://creativecommons.org/licenses/ fluid flows from the glomerulus along the tubule [51]. SGLT2 and by/4.0/). This figure is available as part of a downloadable slideset GLUT2 are responsible for glucose reabsorption in the S1 and S2 Diabetologia (2018) 61:2087–2097 2089

Cellular mechanisms for glucose reabsorption SGLT2 is found in the apical membrane of the S1 and S2 segments of the proximal tubule, while SGLT1 is restricted to the The current models for glucose reabsorption from the glomerular apical membrane of the S3 segment. In rodents, SGLT1 is also filtrate by SGLT2 and SGLT1 in the different segments of the located in the apical membrane of the ascending limb of the loop proximal tubule, S1/S2 and S3, are shown in Fig. 3. In both S1/ of Henle, but the functional significance of this finding is un- S2 and S3 segments, the first stage is glucose transport across the known. We note that currently used immunocytochemical apical membrane by SGLTs. This leads to glucose accumulation methods do not provide quantitative information about the den- within the epithelium, modulated to some extent by intracellular sity or functional activity of targeted membrane proteins. metabolism. The glucose concentration gradient between the cell Actually, it is the number of SGLT proteins and their turnover and plasma in turn drives the second stage: net passive exit of number that determine the functional activity of SGLTs in the cell glucose through the basolateral membrane, towards the plasma, membrane. This information is not available for SGLT2 and via GLUT2. The basolateral Na+/K+ pump (which extrudes three SGLT1 in the apical membrane of S1/S2 and S3 segments. sodium ions for every two potassium ions entering the cell) maintains the sodium gradient across the apical membrane by Functional properties The functional properties of SGLT1 and pumping sodium out of the cell, towards plasma. Inhibition of the SGLT2 have been determined by their expression in heterolo- Na+/K+ pump by cardiac glycosides blocks the pumping of so- gous expression systems such as Escherichia coli, Xenopus dium out of the cell, with the concomitant rise in intracellular laevis oocytes, and cultured cells lacking endogenous activity, sodium concentration. The elimination of the sodium gradient e.g. human embryonic kidney cells (HEK 293) and African across the apical membrane results in the loss of sodium–glucose green monkey kidney, SV40 transformed cells (COS-7) (see [6, cotransport across the apical membrane. Thus, the two-stage pro- 16, 26]). In these systems, the kinetics of sodium–glucose co- cess, together with the absorption of glomerular fluid, accounts transport have been determined as a function of extracellular and for the complete absorption of glucose by the time the filtrate intracellular sodium, sugar and phlorizin concentrations and reaches the end of the proximal tubule (Fig. 2). membrane potential. For now, it suffices to summarise that, at an extracellular NaCl concentration of 150 mmol/l, a membrane potential of −50 mV and at 37°C, the human SGLT2 has an

Cloning renal glucose transporters apparent affinity constant (Km)of5mmol/l(Km =thesubstrate concentration at which the transport velocity is one-half its max- In 1987, members of the Wright laboratory began pioneering imal value), and a sodium:glucose coupling ratio of 1:1. In con- work that resulted in the identification of SGLTs and their func- trast, under the same conditions, human SGLT1 has a glucose tional properties. The rabbit intestinal transporter was first iden- Km of 2 mmol/l and a sodium:glucose coupling ratio of 2:1 (Fig. tified by expression cloning [12], followed by homology cloning 3). These properties are consistent with the hypothesis that, in of the human intestinal SGLT1 and renal SGLT2 transporters humans, the bulk of glucose is absorbed in S1/S2 by SGLT2 and [13, 14]. The SGLTs belong to a human gene family, SLC5, GLUT2, with complete reabsorption occurring in S3, enabled by which contains 12 members, including sodium-coupled trans- the higher affinity of human SGLT1 for glucose and the 2:1 porters for myoinositol, fructose, iodide and short-chain fatty sodium:glucose coupling ratio. acids [6, 15, 16]. SGLT2 (also known as SLC5A2)mRNAis almost exclusively expressed in the kidney, while SGLT1 How do we know that the current models for glucose absorp- (SLC5A1) mRNA is found mainly in the small intestine and is tion in the proximal tubules are correct? Evidence for the only expressed to a small extent in the kidney [6, 17]. A single- importance of SGLT1, SGLT2 and GLUT2 has been obtained cell transcriptomic study of mouse kidney has revealed that Sglt2 using knockout mice and a non-invasive imaging method to is a unique marker gene for cells of the S1 segment of the prox- monitor urinary excretion, micro positron emission tomography imal tubule [18]. Glut2 (also known as Slc2a2) is expressed in (microPET) [27]. Figure 4 shows the time course of excretion of bothS1andS3,whileSglt1 is expressed at a low level along the two PET tracers into the urinary bladder of mice: 2-deoxy-2- 18 proximal tubule, with a somewhat higher level in S3. The SGLT fluoro-[ F]-D-glucose (2-FDG), which is a substrate for genes code for membrane proteins with 14 transmembrane heli- GLUT2 but not for SGLTs, and methyl-4-fluoro-[18F]-4-deoxy- ces, as confirmed by the crystal structures of a bacterial homo- D-glucopyranoside (Me-4FDG), a substrate for SGLTs with only logue, vSGLT [19, 20]. The crystal structures have also provided low affinity for GLUT2. After intravenous injection of 2-FDG, important clues about the SGLT transport mechanism (detailed the tracer is rapidly excreted into the urinary bladder of wild-type below). mice at a rate approximating the filtered load, consistent with the fact that 2-FDG is not a substrate for apical SGLTs. On the other Location of SGLTs in the kidney The localisation of SGLT1 and hand, there is no measurable excretion of Me-4FDG, consistent SGLT2 in the kidney has been determined by immunohisto- with this tracer being a substrate for SGLTs in the apical mem- chemistry using antibodies to the cloned transporters [21–25]. brane and a low-affinity substrate of GLUT2 in the basolateral 2090 Diabetologia (2018) 61:2087–2097

a Lateral intercellular space

Basolateral membrane 1 Na+ SGLT2 Glucose GLUT2 Glucose Km 5 mmol/l

+ Glucose Na :glucose coupling (n)1:1 Glucose Na+

Phlorizin Ki 11 nmol/l Na+/K+ K+ + pump Na K+

Tight junction

b Lateral intercellular space

Basolateral membrane 2 Na+ SGLT1 Glucose GLUT2 Glucose Km 2 mmol/l

+ Glucose Na :glucose coupling (n) 2:1 Glucose Na+

Phlorizin Ki 140 nmol/l Na+/K+ K+ + pump Na K+

Tight junction Fig. 3 Reabsorption of glucose in the proximal tubule. (a) Epithelial cells concentration of 150 mmol/l, a membrane potential of −50 mV and at of the S1 and S2 segments of the proximal tubule express SGLT2 on the 37°C, the human SGLT2 has a Km for glucose of 5 mmol/l, a Ki for apical membrane and GLUT2 on the basolateral membrane. (b) Epithelial phlorizin of 11 nmol/l and a sodium:glucose coupling ratio of 1:1. cells of the S3 segment express SGLT1 on the apical membrane and Under the same conditions, human SGLT1 has a glucose Km of GLUT2 on the basolateral membrane. In both S1/S2 and S3 segments, 2 mmol/l, a phlorizin Ki of 140 nmol/l, and a sodium:glucose coupling glucose reabsorption occurs, first via glucose transport across the apical ratio of 2:1. Adapted from [6], distributed under the terms of the CC BY membrane by SGLTs and then by passive glucose exit towards the plasma 4.0 Attribution License (http://creativecommons.org/licenses/by/4.0/). via GLUT2. The sodium gradient across the apical membrane is main- This figure is available as part of a downloadable slideset tained by the basolateral Na+/K+ pump. At an extracellular NaCl membrane. In GLUT2 knockout mice, Me-4FDG is excreted at a Collectively, this data suggests that, in both mice and men, rate comparable to 2-FDG. While, in SGLT1 and SGLT2 knock- in the early proximal tubule (S1/S2), SGLT2 reabsorbs the out mice, Me-4FDG is excreted, albeit less than the filtered load. bulk of the filtered glucose load and that, in the late proximal Overall, these results are consistent with the roles of SGLT1, tubule (S3), SGLT1 provides a reserve capacity for up to 70% SGLT2 and GLUT2 in renal glucose reabsorption. of the filtered load. In all three segments (S1, S2 and S3), Further evidence comes from measurement of 24 h glucose GLUT2 in the basolateral membrane is essential for complet- excretion in SGLT1, SGLT2 and dual SGLT1/SGLT2 knock- ing glucose absorption across the tubule. out mice [28]. In double knockout mice, the entire filtered glucose load is excreted in the urine, while in single SGLT2 and SGLT1 knockout mice, 67% and 98% of the filtered glu- Inhibition of glucose reabsorption cose load is reabsorbed, respectively. Some patients with truncation mutations in SGLT2 have been found to excrete less External phlorizin acts as a high-affinity, specific, non- than 50% of the filtered glucose load, while those with trun- transported, competitive, dead-end inhibitor of cation mutations in SGLT1 have only mild glucosuria (see: sodium:glucose transport in and out of a cell. It contains a [29]; Table 7.1 in [30]; Table 6 in [6]). Finally, in both mice glucose structure on one end, which binds to the glucose- and humans the ‘knockout’ of functional GLUT2 (truncation binding site on SGLTs, and an aglycone tail (phloretin), which mutations in GLUT2 results in massive glucosuria [9]. binds to the wall of the hydrophilic cavity leading to the Diabetologia (2018) 61:2087–2097 2091

3 within tubules surrounding the glomeruli. This pattern of binding is similar to the hybridisation of Sglt2 (also 2-FDG: WT known as Slc5a2) mRNA in rat kidneys and SGLT2 antibody binding in mouse, rat and human kidneys. We con- 2 Me-4FDG: Glut2−/− clude that dapagliflozin is filtered from the plasma by the glomerulus and then binds to SGLT2 in the early proximal

Me-4FDG: Sglt1−/− tubule, where it inhibits glucose absorption. Biochemical 3

Excretion (MBq) 1 studies in dog kidney have shown that [ H]-phlorizin is filtered by the kidney, binds to the proximal tubule brush

Me-4FDG: Sglt2−/− border membranes, and is displaced and excreted into the urine by excess cold (non-radioactive) phlorizin [34]. 0 Me-4FDG: WT Unlike phlorizin, dapagliflozin and other SGLT2 inhibi- 0102030405060 tors are not excreted into the urine, implying that they are Time (min) reabsorbed further down the nephron and excreted in bile Fig. 4 Urinary excretion of glucose PET tracers, 2-FDG and Me-4FDG − − − − − − [32]. in wild-type, Glut2 / , Sglt1 / and Sglt2 / mice. The total amount of 2- FDG and Me-4FDG in the urinary bladder of representative mice as a function of time after intravenous injection of radiotracer (11 MBq) is shown. The data were fitted to a three-compartmental model for glomer- Molecular mechanisms of glucose transport ular filtration, reabsorption and urinary excretion, showing that the excre- − − by SGLTs and GLUTs tion of 2-FDG in wild-type mice, and Me-4FDG in the Glut2 / mice was equivalent to the filtered glucose load. The excretion of Me-4FDG was greater in Sglt1−/− than Sglt2−/− mice. The entire filtered load of Me- A mechanical model of sodium–glucose cotransport by 4FDG was reabsorbed in wild-type mice. Adapted from [27] distributed SGLTsisshowninFig.6. The vast majority of informa- under the terms of the CC BY-NC 4.0 Attribution License (https:// tion on SGLT-mediated sodium–glucose cotransport has creativecommons.org/licenses/by-nc/4.0/). This figure is available as part of a downloadable slideset come from biochemical and biophysical experiments on human SGLT1 in heterologous expression systems, with a more limited set of data from human SGLT2 experiments. glucose-binding site (described further below). Internal In addition, solving the crystal structure of the bacterial phlorizin, even at high cytoplasmic sodium concentrations, homologue vSGLT and molecular dynamic simulations of is a poor inhibitor of both forward and reverse glucose trans- these structures has provided further insights into the mo- port [31]. lecular mechanisms of SGLTs. We have also learnt from Several studies have been carried out to confirm and the fact that SGLTs belong to a structural family of trans- explain the total inhibition of glucose reabsorption by porters that have a common five-helix inverted repeat phlorizinthatisreportedinhumans[1]. For example, motif, the LeuT structural fold (see [6, 35]). Different using microPET to monitor Me-4FDG excretion into the cotransporters and exchangers have this common structure urinary bladder, we have shown that intravenous injection and so may have similar transport mechanisms. For of phlorizin or the specific SGLT2 inhibitor dapagliflozin cotransporters and exchangers in this family, the substrate rapidly increases excretion of Me-4FDG into the urinary binding sites are located in the middle of the protein, with bladder (Fig. 5a) [32]. Dapagliflozin acts specifically on external and internal gates isolating the substrate from the the kidneys, as shown by microPET imaging of mice extracellular and intracellular solution on each side of the injected with 4-[18F]fluoro-dapagliflozin (F-Dapa) [32]. membrane (Fig. 6). Having determined the structure of F-Dapa binds specifically to the external surface of func- GLUT transporters in the SLC2 gene family [36], it is tional SGLT2 in the plasma membrane (with a binding suggested that glucose transport through these proteins constant of 4 nmol/l) [33] and is displaced by phlorizin has a similar gated mechanism to the cotransporters, but and cold (non-radioactive) dapagliflozin. In rodents, the with the difference that opening and closing of the exter- kidney is the only organ that shows significant specific F- nal and internal gates is not controlled by sodium. Dapa binding (Fig. 5b). As SGLT2 inhibitors only bind to SGLT1 normally couples the inward transport of two sodi- functional SGLT2 proteins in plasma membranes [31], um ions and one glucose molecule in each kinetic cycle. The this extends the immunohistochemical finding of the direction of the transporter is completely reversible. The rate kidney-specific location of SGLT2 in rodents [23]. and direction of transport is simply a function of the extracel- So, where in the kidney does F-Dapa bind? This has lular and intracellular sodium and glucose concentrations and been determined by ex vivo autoradiography (Fig. 5c,d). the polarity and magnitude of the membrane potential. The inhibitor binds to the outer cortex of the whole mouse Cotransport of sodium and glucose by SGLT1 and SGLT2 kidney and (as shown at higher magnification) only generates an electrical current and this provides a biophysical 2092 Diabetologia (2018) 61:2087–2097

a Dapa injection 7 c 6

2 Excretion (MBq) 1

0 0204060 Time (min)

b d 8

SGLT2 density 2

*** 0

Heart Brain Liver Kidney Bladder

Muscle (neck) Harderian gland Submaxillary gland Fig. 5 (a) Time course of Me-4FDG excretion into the urinary bladder of bladder, for which values are presented as percentage ID per total bladder a rat. Me-4FDG (300 MBq) was injected intravenously into rats and volume. Data are presented as the apparent density of SGLT2 in each excretion into the urinary bladder measured using microPET. Its excretion organ, as means + SEM. ***p ≤ 0.001 vs control. The data show that into the urinary bladder (MBq) is plotted as a function of time before and functional SGLT2 is only expressed in the kidney. (c, d) Location of after injection of 1 mg/kg dapagliflozin (Dapa; an SGLT2 inhibitor) at SGLT2 in the mouse kidney, visualised using F-Dapa and ex vivo 20 min. Injection of dapagliflozin causes the rapid excretion of Me-4FDG microautoradiography and H&E staining. A mouse was injected with into the urinary bladder. Similar results were obtained with intravenous 148 MBq F-Dapa and after 15 min the kidney was removed and proc- phlorizin (A. S. Yu and C. Ghezzi, unpublished results). (b) SGLT2 dis- essed. (c) Aligned autoradiogram and H&E-stained image of the whole tribution in a rat, analysed using F-Dapa microPET. Animals were mouse kidney. Scale bar, 1 mm. (d) F-Dapa binding to the tubules sur- injected with F-Dapa (300 MBq) and the distribution of the tracer was rounding a glomerulus. These images show that dapagliflozin is filtered imaged at 60 min using microPET. F-Dapa binding to organs is shown in by glomeruli in the outer renal cortex and then binds to SGLT2 in the control conditions (white bars) or after competition with cold early proximal tubule. Scale bar, 100 μm. Parts (b–d) are adapted with dapagliflozin (black bars). Three-dimensional regions of interest (ROIs) permission of American Society of Nephrology from [32]; permission were drawn over each organ and the data are presented as a percentage of conveyed through Copyright Clearance Center, Inc. This figure is avail- the initial dose per tissue weight (%ID/g), with the exception of the able as part of a downloadable slideset tool to measure the rate of cotransport. The sugar-activated conserved in SGLT2 and other members of the LeuTstructural (inward or outward) current is observed in the presence of family: in SGLT1 the residues involved in coordinating Na+ sodium and is equivalent to the rate of sodium transport. In are A76, I79, S289, S392 and S293. We have evidence that the the case of SGLT1, capacitive currents are generated by volt- Na1 site in human SGLT1 involves residues N78, H83, E102, age jumps in the presence and absence of sodium, in the ab- Y290 and W291 [37]. Sodium binding increases the probabil- sence of sugar. These currents are due to the movement of ity of opening the external gate so that external glucose (and charged or polar amino acid residues of SGLT1 in the mem- phlorizin) can bind. The residues responsible for glucose bind- brane electric field. They have provided a powerful biophys- ing are N78, H83, E102, K321, Q457, Y290 and W291, and ical means of measuring SGLT1 presteady state kinetics and these are conserved in human SGLT2. After binding glucose the number of SGLT1 proteins in the plasma membrane. As (or phlorizin) the outer gate closes to occlude the substrate yet, capacitive currents have not been recorded for human from the external solution and then the inner gate opens to SGLT2. allow glucose and sodium to escape into the cytosol. Finally, In the normal forward transport cycle (Fig. 6), extracellular the inner gate closes and the kinetic cycle continues to the sodium first binds to the Na2 and Na1 sites. The Na2 site is initial starting position. We originally proposed an ordered Diabetologia (2018) 61:2087–2097 2093

OUT

Outer Outer gate vestibule

Glucose

Na1 Inner vestibule Inner Na2 12gate 3

Fig. 6 A mechanical model for sodium-coupled sugar transport. Sodium the inner gate closes to form the inward facing ligand-free conformation (green circle) binds first to the extracellular side (‘OUT’; state 1) to open (state 5). The cycle is completed by the change in conformation to the the outer gate (state 2), permitting the sugar (glucose; yellow hexagon) to outward facing ligand-free (state 1). Phlorizin binds at the second point in bind and be trapped in the bound site (state 3). The binding of both the process (state 2). Adapted from [6], distributed under the terms of the substrates induces a conformational change to an ‘inward facing’ confor- CC BY 4.0 Attribution License (http://creativecommons.org/licenses/by/ mation, resulting in the opening of the inner gate (state 4) and the release 4.0/). This figure is available as part of a downloadable slideset of Na+ and sugar into the cell interior. After the release of both substrates, internal dissociation of glucose and sodium, but recent exper- properties of SGLT2 expressed in oocytes (cultured at iments and molecular dynamic studies favour the simulta- 22°C) were identical to those in HEK 293 cells (cultured neous release of glucose and sodium [38]. One complete cycle at 37°C). The role of MAP17 in SGLT2 expression in takes about 20 ms. oocytes is not yet understood. Although kinetic studies of SGLT2 are not as advanced as those for SGLT1, the transport model is similar (Fig. 6). A major reason for the lack of data on SGLT2 is the Regulation of SGLTs low expression of this transporter in heterologous expression systems, such as X. laevis oocytes and cultured cells In HEK 293T cells, SGLT2 is regulated over the short term by at 22°C. In preliminary kinetic studies on oocytes, the stimulation of protein kinases, Protein kinase A (PKA) and expression level of SGLT2 was less than 1% of that for Protein kinase C (PKC) [43]. This reversible stimulation oc- SGLT1 [39]. On learning from Ana Pajor and Chari Smith curs with a half-time of 10 min and is due to a change in the that raising the temperature to 37°C dramatically in- maximum rate of transport. Insulin mimics the effect of PKA creased SGLT2 activity in cultured cells [40], we were and PKC activation, but this does not occur with deletion of able to perform a robust comparison of the kinetics of the only phosphorylation site on SGLT2, Ser624Ala. PKC and SGLT1andSGLT2inHEK293cells[26, 33]. The major PKA also regulate human SGLT1 in cultured cells, resulting difference between SGLT2 and SGLT1 is that the in rapid changes in the trafficking of an intracellular pool of sodium:glucose coupling ratio is 1:1 for SGLT2 vs 2:1 transporters to the cell membrane. In oocytes, biophysical and for SGLT1 [26]. Other differences include the higher af- electron microscopic studies show that PKA- and PKC- finity of phlorizin to SGLT2 vs SGLT1 (Ki 11 vs induced increases in SGLT1 activity is due to an increase in 140 nmol/l), the higher affinity of gliflozins, such as the number of SGLT1 proteins in the cell membrane, at a rate 7 dapagliflozin to SGLT2 (Ki 4vs400nmol/l),anda of 1 × 10 molecules per second [6, 44]. There was no effect of narrower sugar selectivity of SGLT2 (galactose is a poor insulin on human SGLT1 trafficking. substrate for SGLT2). Michael Coady and his colleagues recently elucidated the reason for low SGLT2 expression in oocytes with the discovery that co-expression of Structure of SGLTs MAP17 was required [41, 42]. Apart from small differences in kinetic variables (likely owing to the difference Following the work of Eric Turk in purifying the bacterial in temperature of culture conditions), the kinetic transporter, vSGLT [45], the x-ray structure was solved in 2094 Diabetologia (2018) 61:2087–2097 two conformations [19, 20]. The structures have provided unique insights into the mechanism of sodium-coupled sugar transport, as incorporated into the transport model (Fig. 6). As aforementioned, the protein has 14 transmembrane helices with a core-inverted repeat of transmembrane helices 1–5and6–10, each containing a discontinuous helix (transmembrane helix [TM]1 and TM6). The substrate binding site is in the middle of the protein, adjacent to the discontinuous helices (N78, H83, E102, K321 Q457, Y290 and W291), and occluded from the external and internal solutions by external and internal gates. The relatively high sequence identities and similar- ities between vSGLT, SGLT1 and SGLT2 have made it possible to construct homology models of the human pro- Fig. 8 The external vestibule of human SGLT1 in the outward facing, teins (Fig. 7). In the sugar occluded state (Fig. 6, state 3) sodium-bound conformation (Fig. 6, state 2). The vestibule was mapped glucose is coordinated by N78, H83, E102, K321, Q457, using fluorescent reagents covalently bound to cysteine residues in the Y290 and W291, and is excluded from contact with the sugar-binding site, e.g. tetramethylrhodamine (TAMRA) bound to external solution by hydrophobic residues (L84, F98 and Y290C. The location of transmembrane helices (TM) of the structural model of SGLT1 are shown (some helices have been removed for clarity), F453) and the internal solution, in part, by Y290. A com- along with the boundary of the 600 Å3 vestibule (blue area) bounded by mon Na2 sodium-binding site in the SGLT1 membrane the outer ends of TM1, TM2, TM3, TM6, TM9 and TM10. Reproduced and other proteins within the LeuT structural family is from [47], distributed under the terms of the Creative Commons proposed at S393 (along with A76, I79, S389 and S392; Attribution-NonCommerical-NoDerivatives International License 4.0 (CC BY-NC-ND; https://creativecommons.org/licenses/by-nc-nd/4.0/). not shown in Fig. 6). Mutations of each of the SGLT1 This figure is available as part of a downloadable slideset sugar and sodium coordination residues dramatically alter sugar binding, whereas mutations of the external gate residues do not [46]. Considerable progress has been made by using the structure of vSGLT to examine the molecular dynamics of SGLTs (interested readers are referred to the literature, e.g. [20, 38]). These and other biophysical studies of human SGLT1 provide insights into the conformational changes associated with sodium binding and isomerisation of the transporter between the outward and inward facing structures (e.g. see [37, 47]).

Inhibitors of glucose reabsorption

The lead compound for the development of SGLT drugs was phlorizin [48]. As indicated above, this plant glucoside is a non-transported, specific competitive inhibitor of SGLT2 and

SGLT1, with a Ki of 11 nmol/l and 140 nmol/l, respectively. It preferentially binds to the external surface of the SGLTs in the presence of external sodium, blocking glucose transport (inward or outward). The aglycones of phlorizin (phloretin) and dapagliflozin are poor, non-competitive inhibitors of SGLTs, indicating that the glycosides bind to both the glucose-binding Fig. 7 Homology model of the human SGLT2 based on the inward facing, occluded conformation of vSGLT (as described in [19]). Helices are pocket and an adjacent lipophilic vestibule leading to the represented as tubes. For clarity, helices −1and11–14 have been removed glucose-binding site (Fig. 8). There is remarkable selectivity and helices 1, 2 and 10 are depicted as transparent. TM3 is coloured in the binding of gliflozins to SGLTs; for example, orange. Highlighted are the residues forming the glucose-binding site dapagliflozin binds to SGLT2 with a much greater affinity and the inner and outer gates EL8a and EL8b are helices in the external loop linking TM7 and TM8. This figure is available as part of a than to SGLT1, and galacto-dapagliflozin has orders of mag- downloadable slideset nitude lower affinity for SGLT1 than SGLT2 [33]. Diabetologia (2018) 61:2087–2097 2095

A mutational analysis of phlorizin binding to SGLT1 Inherited disorders of renal glucose confirms that the sugar-binding site is important for both transporters glucose and phlorizin binding, in that there was a linear relationship between glucose Km and phlorizin Ki [46]. A As previously mentioned, there are three known rare, autoso- major exception is that mutation of an outer gate residue, mal recessive disorders of SGLTs and GLUT2, glucose–ga-

F101C, increased phlorizin Ki by 200-fold with no change lactose malabsorption (OMIM 182380), familial renal in glucose Km. This indicates an interaction, possibly π–π glucosuria (OMIM 233100) and Fanconi–Bickel syndrome bonding, between F101 and the phlorizin aglycone. (OMIM 227810), all of which result in mild to excessive Additional clues about the location of the phlorizin agly- glucosuria (1–150 g [1.73 m]−2 day−1 [see textbox below]). cone binding site comes from voltage clamp fluorometery In glucose–galactose malabsorption, mutations in SGLT1 of SGLT1, which identified that a 600 Å3 vestibule leads to cause a defect in intestinal glucose (and galactose) absorption, the glucose-binding site in the sodium-bound open confor- resulting in mild glucosuria (see [6, 50]). These mutations mation (Fig. 6, state 2, and Fig. 8)[47]. This vestibule is cause malabsorption due to mistrafficking of SGLT1 to the lined with hydrophobic resides on TM1, TM2, TM6, TM9 brush border membrane. In contrast, mutations in SGLT2 andTM10(Fig.7). We anticipate that gliflozins bind to a cause renal glucosuria (from 1 g [1.73 m]−2 day−1 to 150 g similar site in both SGLT1 and SGLT2, and this is support- [1.73 m]−2 day−1) without defects in intestinal absorption. ed by molecular dynamic studies of inhibitor binding to Only in cases of homogeneous truncation mutations in SGLTs [49]. The success of SGLT2 drugs is, in large part, SGLT2 is it reasonable to expect severe glucosuria, but the due to their high affinity for SGLT2 and the limited expres- reserve capacity of SGLT1 should be taken into account. sion of this transporter in the kidney cortex (see above). Unlike glucose–galactose malabsorption, there are no compre- The development of GLUT2 drugs to treat diabetes has not hensive studies of the transport properties of SGLT2 mutants, been practical owing to the close structural and functional largely due to the low expression of SGLT2 in heterologous similarity between GLUT family members and the vital func- expression systems (as detailed above). This, and the incom- tions of GLUTs throughout the body [36]. plete clinical study of individuals with familial renal

Inherited disorders of SGLT1, SGLT2 and GLUT2

Familial renal glucosuria (OMIM 233100) [6, 52] Benign, rare, autosomal recessive disorder Presents as isolated glucosuria (1–150 g [1.73 m]−2 day−1) Homogenous mutations in SGLT2 in 60% of patients Mutations include missense, nonsense, frame shift, splice site and deletion mutations Those with premature stop mutations (e.g. V347X) have severe glucosuria

Glucose–galactose malabsorption (OMIM 182380) [6, 50] Rare autosomal recessive defect in intestinal glucose and galactose absorption Patients have mild renal glucosuria. Newborns (on mother’s milk) present with diarrhoea Homozygous and heterogeneous SGLT1 missense, nonsense, frame shift, splice site and deletion mutations. Mutations cause defects in tracking SGLT1 from endoplasmic reticulum to brush border membrane Therapy is to remove lactose, glucose and galactose from diet

Fanconi–Bickel syndrome (OMIM 227810) [9] A fatal, rare autosomal recessive disorder Characterised by hepatomegaly and glucosuria ranging from 40–150 g [1.73 m]−2 day−1 Missense, nonsense, frame shift and splice site mutations in GLUT2 Those with truncation mutations have severe glucosuria

OMIM, Online Mendelian Inheritance in Man (www.omim.org) 2096 Diabetologia (2018) 61:2087–2097 glucosuria, has led to confusion about the inheritance of this distribution, and reproduction in any medium, provided you give appro- disease. However, we expect familial renal glucosuria to be a priate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. simple autosomal recessive disease. As with SGLT2, truncation mutations of GLUT2 are expected to cause complete excretion of the filtered glucose load if References GLUT2 is the major glucose transporter on the proximal tubule basolateral membrane. However, the phenotype of Fanconi– 1. Chasis H, Jolliffe N, Smith HW (1933) The action of phlorizin on Bickel syndrome is complex owing to the importance of the excretion of glucose, xylose, sucrose, creatinine and urea by GLUT2 in the liver and other organs. man. J Clin Invest 12:1083–1090 Overall, these genetic disorders, caused by mutations in 2. Burg MB, Knepper MA (1986) Single tubule perfusion techniques. – SGLT1, SGLT2 and GLUT2, support animal studies and con- Kidney Int 30:166 170 3. Barfuss DW, Schafer JA (1981) Differences in active and passive firm the importance of the three affected proteins in glucose transport along the proximal nephron. Am J Phys 241: reabsorbing the filtered glucose load in healthy humans. F322–F332 4. Schultz SG, Curran PF (1970) Coupled transport of sodium and organic solutes. Physiol Rev 50:637–718 5. Hopfer U, Nelson K, Perrotto J, Isselbacher KJ (1973) Glucose Summary transport in isolated brush border membrane from rat small intestine. J Biol Chem 248:25–32 Thirty years ago, the cloning of human SGLT1, SGLT2 and 6. Wright EM, Loo DD, Hirayama BA (2011) Biology of human GLUT2 was a major breakthrough in the identification of trans- sodium glucose transporters. Physiol Rev 91:733–794 porters involved in renal glucose absorption. Subsequent ad- 7. Fanconi G, Bickel H (1949) Chronic aminoaciduria (amino acid diabetes or nephrotic-glucosuric dwarfism) in glycogen storage vances in molecular biology permitted the development of ge- and cystine disease. Helv Paediatr Acta 4:359–396 [article in netic mouse models lacking SGLT1, SGLT2 and GLUT2, fur- German] thering our understanding of the role of these transporters in the 8. Santer R, Schneppenheim R, Dombrowski A, Gotze H, Steinmann proximal tubule. Based on the assumption that SGLT2 was re- B, Schaub J (1997) Mutations in GLUT2, the gene for the liver-type glucose transporter, in patients with Fanconi-Bickel syndrome. Nat sponsible for the bulk of glucose reabsorption, the pharmaceuti- Genet 17:324–326 cal industry in Japan, USA and Europe invented SGLT2 inhibi- 9. Santer R, Groth S, Kinner M et al (2002) The mutation spectrum of tors to treat type 2 diabetes. Although these drugs do lower blood the facilitative glucose transporter gene SLC2A2 (GLUT2) in pa- glucose levels in patients with diabetes without causing tients with Fanconi-Bickel syndrome. Hum Genet 110:21–29 hypoglycaemia, they are only capable of a 50% inhibition of 10. Mudaliar S, Polidori D, Zambrowicz B, Henry RR (2015) Sodium- glucose cotransporter inhibitors: effects on renal and intestinal glu- glucose reabsorption and are, thus, not so effective as a mono- cose transport: from bench to bedside. Diabetes Care 38:2344–2353 therapy. In retrospect, given the huge reserve capacity of SGLT1 11. Gallo LA, Wright EM, Vallon V (2015) Probing SGLT2 as a ther- in the late proximal tubule, one can argue that dual SGLT2 and apeutic target for diabetes: basic physiology and consequences. – SGLT1 drugs may be more effective in controlling renal glucose Diab Vasc Dis Res 12:78 89 12. Hediger MA, Coady MJ, Ikeda TS, Wright EM (1987) Expression excretion. cloning and cDNA sequencing of the Na+/glucose co-transporter. Nature 330:379–381 Acknowledgements We are grateful to all members of the Wright lab and 13. Hediger MA, Turk E, Wright EM (1989) Homology of the human our colleagues over the past 30 years for all their contributions to SGLT intestinal Na+/glucose and Escherichia coli Na+/proline biology. J. Park (University of Pennsylvania, Philadelphia, PA, USA) cotransporters. Proc Natl Acad Sci U S A 86:5748–5752 provided expert guidance in the analysis of gene expression in the mouse 14. Wells RG, Pajor AM, Kanai Y, Turk E, Wright EM, Hediger MA kidney. (1992) Cloning of a human kidney cDNA with similarity to the sodium-glucose cotransporter. Am J Phys 263:F459–F465 Funding Our research has been funded by grants from the National 15. Wright EM, Turk E (2004) The sodium/glucose cotransport family Institutes of Health (DK19567 and DK07713). SLC5. Pflugers Arch 447:510–518 16. Wright EM (2013) Glucose transport families SLC5 and SLC50. Duality of interest The authors declare that there is no duality of interest Mol Asp Med 34:183–196 associated with this manuscript. 17. Nishimura M, Naito S (2005) Tissue-specific mRNA expression profiles of human ATP-binding cassette and solute carrier transport- Contribution statement All authors were responsible for drafting the er superfamilies. Drug Metab Pharmacokinet 20:452–477 article and revising it critically for important intellectual content, and have 18. Park J, Shrestha R, Qiu C et al (2018) Single-cell transcriptomics of approved the version to be published. the mouse kidney reveals potential cellular targets of kidney disease. Science 360:758–763 19. Faham S, Watanabe A, Besserer GM et al (2008) The crystal structure of a sodium galactose transporter reveals mechanistic insights into Na+/sugar symport. Science 321:810–814 Open Access This article is distributed under the terms of the Creative 20. Watanabe A, Choe S, Chaptal V et al (2010) The mechanism of Commons Attribution 4.0 International License (http:// sodium and substrate release from the binding pocket of vSGLT. creativecommons.org/licenses/by/4.0/), which permits unrestricted use, Nature 468:988–991 Diabetologia (2018) 61:2087–2097 2097

21. Cramer SC, Pardridge WM, Hirayama BA, Wright EM (1992) 37. Loo DD, Jiang X, Gorraitz E, Hirayama BA, Wright EM (2013) Colocalization of GLUT2 glucose transporter, sodium/glucose Functional identification and characterization of sodium binding cotransporter, and gamma-glutamyl transpeptidase in rat kidney sites in Na symporters. Proc Natl Acad Sci U S A 110:E4557– with double-peroxidase immunocytochemistry. Diabetes 41:766– E4566 770 38. Adelman JL, Ghezzi C, Bisignano P et al (2016) Stochastic steps in 22. Balen D, Ljubojevic M, Breljak D et al (2008) Revised immunolo- secondary active sugar transport. Proc Natl Acad Sci U S A 113: calization of the Na+-D-glucose cotransporter SGLT1 in rat organs E3960–E3966 with an improved antibody. Am J Phys Cell Phys 295:C475–C489 39. Kanai Y, Lee WS, You G, Brown D, Hediger MA (1994) The 23. Sabolic I, Vrhovac I, Eror DB et al (2012) Expression of Na+−D- human kidney low affinity Na+/glucose cotransporter SGLT2. glucose cotransporter SGLT2 in rodents is kidney-specific and ex- Delineation of the major renal reabsorptive mechanism for D-glu- hibits sex and species differences. Am J Phys Cell Phys 302: cose. J Clin Invest 93:397–404 C1174–C1188 40. Pajor AM, Randolph KM, Kerner SA, Smith CD (2008) Inhibitor 24. Vrhovac I, Balen Eror D, Klessen D et al (2015) Localizations of binding in the human renal low- and high-affinity Na+/glucose Na(+)-D-glucose cotransporters SGLT1 and SGLT2 in human kid- cotransporters. J Pharmacol Exp Ther 324:985–991 ney and of SGLT1 in human small intestine, liver, lung, and heart. 41. Coady MJ, El Tarazi A, Santer R et al (2017) MAP17 is a necessary Pflugers Arch 467:1881–1898 activator of renal Na+/glucose cotransporter SGLT2. J Am Soc 25. Scafoglio C, Hirayama BA, Kepe Vet al (2015) Functional expres- Nephrol 28:85–93 sion of sodium-glucose transporters in cancer. Proc Natl Acad Sci U 42. Coady MJ, Wallendorff B, Lapointe JY (2017) Characterization of SA112:E4111–E4119 the transport activity of SGLT2/MAP17, the renal low-affinity Na+/ 26. Hummel CS, Lu C, Loo DD, Hirayama BA, Voss AA, Wright EM glucose cotransporter. Am J Physiol Ren Physiol 313:F467–F474 (2011) Glucose transport by human renal Na+/D-glucose 43. Ghezzi C, Wright EM (2012) Regulation of the human Na+−de- cotransporters SGLT1 and SGLT2. Am J Phys Cell Phys 300: pendent glucose cotransporter hSGLT2. Am J Phys Cell Phys 303: C14–C21 C348–C354 27. Sala-Rabanal M, Hirayama BA, Ghezzi C et al (2016) Revisiting 44. Hirsch JR, Loo DD, Wright EM (1996) Regulation of Na+/glucose the physiological roles of SGLTs and GLUTs using positron emis- cotransporter expression by protein kinases in Xenopus laevis oo- sion tomography in mice. J Physiol 594:4425–4438 cytes. J Biol Chem 271:14740–14746 28. Powell DR, DaCosta CM, Gay J et al (2013) Improved glycemic 45. Turk E, Kim O, le Coutre J et al (2000) Molecular characterization control in mice lacking Sglt1 and Sglt2. Am J Physiol Endocrinol of Vibrio parahaemolyticus vSGLT: a model for sodium-coupled Metab 304:E117–E130 sugar cotransporters. J Biol Chem 275:25711–25716 29. Santer R, Kinner M, Lassen CL et al (2003) Molecular analysis of 46. Sala-Rabanal M, Hirayama BA, Loo DD, Chaptal V, Abramson J, the SGLT2 gene in patients with renal glucosuria. J Am Soc Wright EM (2012) Bridging the gap between structure and kinetics Nephrol 14:2873–2882 of human SGLT1. Am J Phys Cell Phys 302:C1293–C1305 30. Wright E (2009) Diseases of renal glucose handling. In: Lifton R, 47. Gorraitz E, Hirayama BA, Paz A, Wright EM, Loo DDF (2017) Somlo S, Giebisch GH, Seldin DW (eds) Genetic diseases of the Active site voltage clamp fluorometry of the sodium glucose kidney. Elsevier, New York, pp 131–140 cotransporter hSGLT1. Proc Natl Acad Sci U S A 114:E9980– 31. Ghezzi C, Hirayama BA, Gorraitz E, Loo DD, Liang Y,Wright EM E9988 (2014) SGLT2 inhibitors act from the extracellular surface of the 48. Ehrenkranz JR, Lewis NG, Kahn CR, Roth J (2005) Phlorizin: a cell membrane. Phys Rep 2:e12058 review. Diabetes Metab Res Rev 21:31–38 32. Ghezzi C, Yu AS, Hirayama BA et al (2017) Dapagliflozin binds 49. Bisignano P, Kalyanaraman C, Ghezzi C et al (2017) Structural specifically to sodium-glucose cotransporter 2 in the proximal renal insights into sodium-dependent sugar transporters and their inhibi- tubule. J Am Soc Nephrol 28:802–810 tion mechanism. Biophys J 128a:112 (abstract) 33. Hummel C, Lu C, Liu J et al (2012) Structural selectivity of human 50. Wright EM, Martin MG, Turk E (2001) Familial glucose-galactose SGLT inhibitors. Am J Phys Cell Phys 302:C373–C382 malabsorption and hereditary renal glycosuria. In: Scriver CR, 34. Silverman M (1974) The in vivo localization of high-affinity Beaudet AL, Sly WS et al (eds) The metabolic and molecular bases phlorizin receptors to the brush border surface of the proximal of inherited disease. McGraw-Hill, New York, pp 4891–4908 tubule in dog kidney. Biochim Biophys Acta 339:92–102 51. Frohnert PP, Hohmann B, Zwiebel R, Baumann K (1970) Free flow 35. Abramson J, Wright EM (2009) Structure and function of Na(+)- micropuncture studies of glucose transport in the rat nephron. symporters with inverted repeats. Curr Opin Struct Biol 19:425– Pflugers Arch 315:66–85 432 52. Santer R, Calado J (2010) Familial renal glucosuria and SGLT2: 36. Yan N (2015) Structural biology of the major facilitator superfamily from a mendelian trait to a therapeutic target. Clin J Am Soc transporters. Annu Rev Biophys 44:257–283 Nephrol 5:133–141